Prices stabilization for inexact unit-commitment problems (Q2636928)

This paper gives useful results with respect to mathematical economy as well as to duality in optimization theory. The authors prove and discuss new ideas to solve special difficulties which arise, if one has to solve (big) commitment problems. They also test the found results using real-life examples. Dealing with the problems they of course use constraint dualization and as the mentioned difficulties they especially explain mismatch of demand and bad behaviour of approximate solutions (prices) of dual problems. To improve known solution methods they introduce successfully penalization and stabilization techniques. It is interesting, that Fenchel conjugation is used proving results varying between primal and dual programs, also, that perturbations, which arise as a result of penalization and stabilization in the considered programs do not (or hardly) have influence f. i. to recover exact prices. As a last remark we mention a hint of the authors that the assumptions on the oracle noise are inspired from noise removal in image processing, and they refer to \textit{D. Strong} and \textit{T. Chan} [Inverse Probl. 19, No. 6, S165--S187 (2003; Zbl 1043.94512)].